Design Of Driving Behavior Decision And MPC Controller For Intelligent Connected Vehicle

The advantages of intelligent vehicle in energy consumption,efficiency,safety and convenience make it the future direction of vehicle development.The "networking" of intelligent connected vehicle enables it to obtain the information of other vehicles through the road test equipment,thus increasing the accuracy of intelligent connected vehicle decision-making.This thesis,based on intelligent connected vehicle,designs an intelligent connected vehicle decision algorithm based on decision tree,and designs controllers for cruise control,trajectory tracking and trajectory planning based on Model Predictive Control(MPC).The main research contents are as follows:(1)Develop collision avoidance decision algorithm for vehicles in typical traffic scenarios.Using abstraction processing of the typical traffic scenarios.Constructing a safety distance model for vehicle collision.Presenting a vehicle collision avoidance decision-making based on ID3 decision tree algorithm through establishing data samples conforms to the traffic rules and driving habits.Not only using the vehicle perception layer to collecte environmental information,but also considering the traffic information providing by V2 X road test equipment.It is conductive to the stable application of the algorithm in the dynamic traffic scenarios of the internet of vehicles.(2)Building the vehicle model and design intelligent connected vehicle controller.The dynamics vehicle model with three degrees of freedom and tire model were built,and the constraint of vehicle side deflection angle was given.On this basis,the adaptive cruise controller(ACC)based on MPC was designed,and the constraints of the distance between cars,driving speed and acceleration were set.A lane change controller based on MPC is designed.The geometric curve of lane change superimposed by constant vel ocity offset curve and sinusoidal curve of lane change is selected as the planned curve for lane change.A trajectory tracking controller with constraints of the side deflection angle of vehicle centroid,the side deflection angle of tire and acceleration is designed.the MPC collision avoidance trajectory planning controller was designed for comparison.Particle swarm optimization(pso)was used to optimize the time domain parameters of MPC at different vehicle speeds.(3)Four kinds of simulation verification were set up for the designed controller.Four vehicle operating conditions after the decision were simulated and verified on the Carsim/Simulink simulation platform.The constant speed cruising mode adopts the PID speed following controller.The simulation result shows that the automatic vehicle can track the cruise speed and accelerate well.ACC controller is adopted in the steady-state following condition,and the vehicle can respond to the change of the motion state of the preceding vehicle in time.Trajectory planning controller and trajectory tracking controller are adopted for lane change condition,in which the geometric curve of lane change is planned for the lane change-cruise condition,and the vehicle can track up lane change curve very well.For lane change-overtaking conditions,geometric method was used to plan lane change overtaking and MPC collision avoidance trajectory planning for comparative simulation.The simulation result shows that the MPC track planner can avoid the preceding vehicle very well and successfully change the lane to overtake.In this thesis,a complete driving decision and motion control system was designed for intelligent connected vehicles in typical traffic scenarios,and the controller was simulated and verified under typical vehicle operating conditions through carsim/simulink co-simulation platform.The research results can be used for reference in the research of decision and control of intelligent connected vehicle.